Facial Emotion Recognition Predicts Alexithymia Using Machine Learning

Abstract

Objective. Alexithymia, as a fundamental notion in the diagnosis of psychiatric disorders, is characterized by deficits in emotional processing and, consequently, difficulties in emotion recognition. Traditional tools for assessing alexithymia, which include interviews and self-report measures, have led to inconsistent results due to some limitations as insufficient insight. Therefore, the purpose of the present study was to propose a new screening tool that utilizes machine learning models based on the scores of facial emotion recognition task. Method. In a cross-sectional study, 55 students of the University of Tabriz were selected based on the inclusion and exclusion criteria and their scores in the Toronto Alexithymia Scale (TAS-20). Then, they completed the somatization subscale of Symptom Checklist-90 Revised (SCL-90-R), Beck Anxiety Inventory (BAI) and Beck Depression Inventory-II (BDI-II), and the facial emotion recognition (FER) task. Afterwards, support vector machine (SVM) and feedforward neural network (FNN) classifiers were implemented using K-fold cross validation to predict alexithymia, and the model performance was assessed with the area under the curve (AUC), accuracy, sensitivity, specificity, and F1-measure. Results. The models yielded an accuracy range of 72.7–81.8% after feature selection and optimization. Our results suggested that ML models were able to accurately distinguish alexithymia and determine the most informative items for predicting alexithymia. Conclusion. Our results show that machine learning models using FER task, SCL-90-R, BDI-II, and BAI could successfully diagnose alexithymia and also represent the most influential factors of predicting it and can be used as a clinical instrument to help clinicians in diagnosis process and earlier detection of the disorder.

Emotion Processing in Peripheral Neuropathic Pain: An Observational Study

“…In line with previous studies [ 6 , 46 , 47 ], we found reduced performance on the overall EK-60F in alexithymic patients. In these patients, left CDT z-scores also correlated with the EK-60F subscore for surprise, which is a predictor of alexithymia [ 47 ]. In contrast, in non-alexithymic patients, sensitivity to heat pain is inversely related to the EK-60F subscore for fear.…”

Machine learning‐based classifying of risk‐takers and risk‐aversive individuals using resting‐state EEG data: A pilot feasibility study

“…The classification results have been evaluated using four metrics: accuracy (%), sensitivity (%), specificity (%), and F 1‐measure (%) scores. Given the basic statistics of true positive (TP) represents the number of risk‐taking managers detected correctly, true negative (TN) measures the number of risk‐averse managers predicted correctly, false negative (FN) represents the number of risk‐taking managers detected as risk‐averse managers, and false positive (FP) represents the number of risk‐averse managers predicted as risk‐taking managers (Farhoumandi et al., 2021). Therefore, the evaluation metrics can be calculated as follows: $$$$\begin{equation}{\mathrm{Accuracy\ }} = \ \frac{{{\mathrm{TP}} + {\mathrm{TN}}}}{{{\mathrm{TP}} + {\mathrm{TN}} + {\mathrm{FP}} + {\mathrm{FN\ }}}}\end{equation}$$$$ $$$$\begin{equation}{\mathrm{Sensitivity\ }} = \frac{{{\mathrm{TP\ }}}}{{{\mathrm{TP}} + {\mathrm{FN}}}}\end{equation}$$$$ $$$$\begin{equation}{\mathrm{Specifcity\ }} = \frac{{{\mathrm{TN\ }}}}{{{\mathrm{TN}} + {\mathrm{FP}}}}\end{equation}$$$$ $$$$\begin{equation}{\mathrm{Precision\ }} = \frac{{{\mathrm{TP\ }}}}{{{\mathrm{TP}} + {\mathrm{FP}}}}\end{equation}$$$$ $$$$\begin{equation}F1 - {\mathrm{measure\ }} = \ 2 \times \ \frac{{{\mathrm{precision}} \times {\mathrm{sensitivity}}}}{{{\mathrm{precision}} + {\mathrm{sensitivity}}}} = \frac{{2{\mathrm{TP}}}}{{2{\mathrm{TP}} + {\mathrm{FP}} + {\mathrm{FN}}}}\end{equation}$$$$…”

“…The classification results have been evaluated using four metrics: accuracy (%), sensitivity (%), specificity (%), and F 1‐measure (%) scores. Given the basic statistics of true positive (TP) represents the number of risk‐taking managers detected correctly, true negative (TN) measures the number of risk‐averse managers predicted correctly, false negative (FN) represents the number of risk‐taking managers detected as risk‐averse managers, and false positive (FP) represents the number of risk‐averse managers predicted as risk‐taking managers (Farhoumandi et al., 2021 ). Therefore, the evaluation metrics can be calculated as follows: …”

Techniques for facial affective computing: A review

“…The result shows that the CNN-BiLSTM was 98.75% accurate. Farhoumandi et al (2021) tried to predict alexithymia using a facial emotion recognition model. Several students from the University of Tabriz were selected based Stud.…”

“…We have also observed that many researchers have employed machine learning and deep learning techniques for emotion detection purposes. As reviewed in this research, one of the algorithms that were being used is the SVM which was used by researchers like (Gupta 2018;Farhoumandi et al, 2021;Siam et al, 2022) went ahead and compared the SVM with other algorithms which include KNN, NB, LR, RF and a multilayer perceptron (MLP) algorithm, eventually the MLP outperform the compared machine learning algorithm. Other algorithms used are deep learning algorithms which include the CNN algorithm, which was used by (Punidha et al, 2020;Singh et al2022).…”